July 2010Xianming Shi Yajun LiuMatthew Mooney Michael Berry Barrett HubbardLaura FayAndrea Beth Leonard

WA-RD 741.1

Office of Research & Library Services

WSDOT Research Report

Effect of Chloride-based Deicers on Reinforced Concrete Structures

i

E F F E C T OF C H L OR I DE -B A SE D DE I C E R S ON R E I NF OR C E D C ONC R E T E ST R UC T UR E S

Final Report

Prepared for the

by

Xianming Shi, Ph.D., P.E. (Principal Investigator)

Yajun Liu, Ph.D.

Matthew Mooney

Michael Berry, Ph.D.

Barrett Hubbard

Laura Fay, M.Sc.

Andrea Beth Leonard

Corrosion & Sustainable Infrastructure Laboratory

Western Transportation Institute

Montana State University, Bozeman, MT 59717

Last revised: July 15, 2010

ii

DISCLAIMER

The contents of this report reflect the views of the authors, who are responsible for the facts and the accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the Washington State Department of Transportation or the Federal Highway Administration. This report does not constitute a standard, specification, or regulation.

Reference herein to any specific commercial products, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the authors or the project sponsors.

Alternative accessible formats of this document will be provided upon request. Persons with disabilities who need an alternative accessible format of this information, or who require some other reasonable accommodation to participate, should contact Catherine Heidkamp, Assistant Director for Communications and Information Systems, Western Transportation Institute, Montana State University, PO Box 174250, Bozeman, MT 59717-4250, telephone number 406-994-7018, e-mail: KateL@coe.montana.edu.

.

ACKNOWLEDGEMENTS

The authors acknowledge the financial support provided by the Washington State Department of Transportation as well as the Research & Innovative Technology Administration (RITA) at the U.S. Department of Transportation for this project. The authors would like to thank the WSDOT Research Manager Kim Willoughby and the technical panel consisting of Rico Baroga, Tom Root, DeWayne Wilson, Linda M. Pierce, Jeff S. Uhlmeyer, and Douglas Pierce, for providing the continued support throughout this project. We owe our thanks to Felix Chandra, Mike Weeks, and Greg Mckinnon at the Stoneway Concrete Inc. (Renton, WA) for assisting with the fabrication and quality assurance of all the concrete specimens used in this study. We appreciate the following professionals who provided assistance to this research: Jim Weston (WSDOT), Mary Gilmore (WSDOT), and Mark Peterson (Montana Department of Transportation). We also thank our collaborators at the Montana State University (MSU): Dr. Recep Avci of the Imaging and Chemical Analysis Laboratory for the use of FESEM/EDX instrumentation; Dr. Ed E. Adams of the Civil Engineering Department for the use of environmental chamber. Finally, we owe our thanks to the following individuals at the Western Transportation Institute for providing help in various stages of the laboratory investigation: Marijean M. Peterson, Dr. Tuan Anh Nguyen, Dr. Tongyan Pan, Dr. Wei Chu, Steven Anderson, Justin P. Hauck, Andrew Gilbert, Brian Herting, Eric Schon, and Tristan J. Dunlap.

iii

1. REPORT NO. 2. GOVERNMENT ACCESSION NO. 3. RECIPIENTS CATALOG NO WA-RD 741.1

4. TITLE AND SUBTITLE 5. REPORT DATE Effect of Chloride-based Deicers on Reinforced Concrete Structures

July 2010 6. PERFORMING ORGANIZATION CODE

7. AUTHOR(S) 8. PERFORMING ORGANIZATION REPORT NO. Xianming Shi, Yajun Liu, Matthew Mooney, Michael Berry, Barrett Hubbard, Laura Fay, Andrea Beth Leonard

9. PERFORMING ORGANIZATION NAME AND ADDRESS 10. WORK UNIT NO. Western Transportation Institute PO Box 174250, Montana State University Bozeman, MT 59717-4250

11. CONTRACT OR GRANT NO.

12. CO-SPONSORING AGENCY NAME AND ADDRESS 13. TYPE OF REPORT AND PERIOD COVERED Washington State Department of Transportation - Research 310 Maple Park Avenue SE PO Box 47372, Olympia WA 98504-7372 Research Manager: Kim Willoughby 360-705-7978

Final Report, Dec. 1995-July 2010 14. SPONSORING AGENCY CODE

15. SUPPLEMENTARY NOTES Conducted in cooperation with the U.S. Department of Transportation, Federal Highway Administration. 16. ABSTRACT We conducted an extensive literature review and performed laboratory tests to assess the effect of chloride-based deicers on the rebars and dowel bars in concrete and to determine whether or not deicer corrosion inhibitors help preserve the transportation infrastructure. The laboratory investigation exposed concrete samples to four common chloride-based deicers for approximately one year or less, for natural diffusion at room temperature or for cyclic exposure with wet/dry and temperature cycling. Under the experimental conditions in this study, the corrosion inhibitors in deicers helped to preserve the strength of concrete undergoing temperature and wet/dry cycles. While they also slowed down the chloride ingress and subsequent corrosion initiation of steel in concrete, such benefits seem to diminish once the active corrosion of the rebar is initiated. There were small differences in the corrosion behavior of various dowel bars investigated, limited by the short duration of this study. Corrosion inhibitors and other additives in deicers did not show significant benefit in inhibiting the chemical changes of concrete induced by cations and/or anions in deicers. Agencies should be aware of the deleterious effects magnesium chloride deicers can pose on the concrete strength, even though the inhibited magnesium chloride deicers can pose less corrosion risk for steel in concrete, relative to sodium-chloride-based deicers with or without inhibitor. To anticipate better return on investment regarding the preservation of reinforced concrete, agencies should focus on improved concrete mix designs with less permeability instead of procuring the more costly (inhibited) deicers. 17. KEY WORDS 18. DISTRIBUTION STATEMENT Winter maintenance, snow and ice control, deicing chemicals, rebar corrosion, dowel bar corrosion, reinforced concrete, chloride ingress, sodium chloride, magnesium chloride, calcium chloride

No restrictions. This document is available to the public through the National Technical Information Service, Springfield, VA 22161; www.ntis.gov

19. SECURITY CLASSIF. (of this report) 20. SECURITY CLASSIF. (of this page) 21. NO. OF PAGES 22. PRICE None None

iv

This page was intentionally left blank.

v

Table of Contents

ACKNOWLEDGEMENTS ......................................................................................................... II

LIST OF TABLES ......................................................................................................................... X

ABBREVIATIONS AND ACRONYMS ................................................................................... XI

E X E C UT I V E SUM M A R Y ........................................................................................................1

C H A PT E R 1. I NT R ODUC T I ON ...............................................................................................1

1.1. PROBLEM STATEMENT ...........................................................................................5

1.2. BACKGROUND .................................................................................................5

1.2.1. Chemicals for Winter Road Maintenance .......................................................5 1.2.2. Corrosion of Steel in Concrete and Concrete Permeability ............................7 1.2.3. Mechanisms of Steel Corrosion in Concrete ...................................................8 1.2.4. Transport of Chlorides and Inhibitors in Concrete: The State of

Knowledge ..................................................................................................9 1.2.5. Deicer Impact on Concrete Performance and Durability ...............................9 1.2.6. Deicer Corrosion to Metals in Transportation Infrastructure ........................9

1.3. STUDY OBJECTIVES ..............................................................................................10

1.4. HOW THIS REPORT IS ORGANIZED .....................................................................11

1.5. REFERENCES .........................................................................................................11

C H A PT E R 2. M E T H ODOL OG Y ............................................................................................13

2.1. EXPERIMENTAL ....................................................................................................13

2.1.1. Deicers of Interest ........................................................................................ 13 2.1.2. Design of Deicing Ponding Experiments ..................................................... 15 2.1.3. Concrete Samples: Materials and Preparation ............................................ 16 2.1.4. Chloride Sensors .......................................................................................... 21 2.1.5. Deicing Ponding Experiments ...................................................................... 23 2.1.6. Compression Testing of Concrete Core Samples ......................................... 29 2.1.7. FESEM/EDX Measurements ........................................................................ 29 2.1.8. Estimating the Apparent Chloride Diffusion Coefficient Using the EDX

Data ......................................................................................................... 30 2.2. PREDICTIVE MODELS BASED ON ARTIFICIAL NEURAL NETWORKS ..................31

2.3. REFERENCES .........................................................................................................33

vi

C H A PT E R 3. L A B OR A T OR Y I NV E ST I G A T I ON A ND NE UR A L NE T W OR K S M ODE L I NG

OF DE I C E R I NG R E SS I NT O POR T L A ND C E M E NT C ONC R E T E A ND I T S

C OR R OSI ON I M PL I C A T I ONS .........................................................................34

3.1. CHLORIDE DIFFUSION COEFFICIENT OF DEICERS IN CONCRETE ......................34

3.2. EFFECT OF DEICER PONDING EXPERIMENTS ON THE CONCRETE

STRENGTH ....................................................................................................36

3.2. DATA ANALYSIS FOR DEICER PONDING EXPERIMENTS AND RELATED

ANN-BASED PREDICTIVE MODELS .............................................................38

3.2.1. ANN Model for the Chloride Sensor Potential ............................................. 40 3.2.2. ANN Model for the Top Bar OCP ................................................................ 46 3.2.3. ANN Model for the Macro-cell Current ....................................................... 53 3.2.4. ANN Model for the Top Bar Corrosion Rate ............................................... 55

3.3. CONCLUSIONS .......................................................................................................64

3.4. REFERENCES .........................................................................................................66

CHAPTER 4. CHEMICAL CHANGE OF PORTLAND CEMENT CONCRETE EXPOSED TO

DEICERS ........................................................................................................67

4.1. INTRODUCTION .....................................................................................................67

4.2. RESULTS AND DISCUSSION ...................................................................................68

4.2.1. Effect of Non-Inhibited NaCl on the Chemistry of PCC .............................. 69 4.2.2. Effect of Inhibited NaCl on the Chemistry of PCC ...................................... 73 4.2.3. Effect of the Inhibited CaCl2 Deicer on the Chemistry of PCC ................... 75 4.2.4. Effect of the Inhibited MgCl2 Deicer on the Chemistry of PCC .................. 78

4.3. CONCLUSIONS .......................................................................................................80

4.4. REFERENCES .........................................................................................................80

C H A PT E R 5. C ONC L USI ONS, R E SE A R C H NE E DS A ND R E C OM M E NDA T I ONS F OR

I M PL E M E NT A T I ON ........................................................................................83

5.1. CONCLUSIONS .......................................................................................................83

5.2. RESEARCH NEEDS .................................................................................................83

5.3. RECOMMENDATIONS FOR IMPLEMENTATION .....................................................86

vii

List of Figures

Figure 1. Relative corrosivity of the four deicers used in this study, based on the PNS/NACE method. ....................................................................................... 14

Figure 2. (a) A custom-made wooden mold used to fabricate the reinforced pavement samples; (b) a portion of the mold to make the pond; (c) a fabricated concrete sample with the ponding mold not yet removed .............................. 18

Figure 3. Some reinforced pavement samples (a) during and (b) after the fabrication process (including pouring and vibration). ..................................................... 19

Figure 4. Quality control of fabricated concrete samples: (a) slump test; (b) air content test; and (c) cylinder molds for fabricating compressive strength test specimens. ................................................................................................ 20

Figure 5. Custom-made chloride sensors (dark color) welded to copper wires (red color), before being embedded in the concrete samples. ................................ 21

Figure 6. Electrochemical potential of custom-made chloride sensors as a function of chloride concentration in (a) neutral solutions; (b) highly-alkaline simulated concrete pore solutions. .................................................................. 22

Figure 7. Electrochemical potential of custom-made chloride sensors as a function of NaCl concentration in 28-day concrete samples. ............................................ 23

Figure 8. (a) Schematics and (b) experimental setup for the pressurized transport of deicers into concrete ....................................................................................... 24

Figure 9. (a) The environmental chamber used for this study; (b) The reinforced concrete samples for the Cyclic Exposure Test on the shelves inside the chamber. .......................................................................................................... 25

Figure 10. Temperature and wet/dry cycles experienced by the reinforced concrete samples during the Cyclic Exposure Test ....................................................... 26

Figure 11. Time-dependent behavior of macro-cell current [2] ....................................... 27

Figure 12. Equivalent electric circuit used in the EIS data analysis ................................. 29

Figure 13. Typical multi-layer feed-forward neural network architecture ....................... 32

Figure 14. Chloride diffusion coefficient as a function of deicer type and mix design .... 35

Figure 15. A typical predicted chloride profile in concrete .............................................. 36

Figure 16. Average compressive strength of concrete samples as a function of deicer type and mix design: after the Natural Diffusion Test .................................... 37

Figure 17. Average compressive strength of concrete samples as a function of deicer type and mix design: after the Cyclic Exposure Test ..................................... 38

Figure 18. Relationship between experimental and modeled chloride sensor potential ... 41

Figure 19. Predicted chloride sensor reading as a function of deicer type, test type and mix design ....................................................................................................... 42

viii

Figure 20. Predicted response surface of chloride sensor reading in pavement mix as a function of: (a) exposure duration and wet percent time; (b) reading temperature and daily average exposure temperature ..................................... 44

Figure 21. Predicted response surface of chloride sensor reading in bridge mix as a function of: (a) exposure duration and wet percent time; (b) reading temperature and daily average exposure temperature ..................................... 45

Figure 22. Relationship between experimental and modeled top bar OCP ...................... 46

Figure 23. Predicted top bar OCP as a function of deicer type, test type and bar type in the (a) bridge mix; (b) pavement mix ......................................................... 48

Figure 24. Predicted response surface of OCP of epoxy coated dowel bar in pavement mix as a function of: (a) daily average Cl- concentration and wet percent time; (b) reading temperature and daily average exposure temperature ......... 50

Figure 25. Predicted response surface of OCP of uncoated rebar in bridge mix as a function of: (a) daily average Cl- concentration and wet percent time; (b) reading temperature and daily average exposure temperature ........................ 52

Figure 26. Relationship between experimental and modeled macro-cell current ............. 54

Figure 27. Predicted macro-cell current as a function of deicer type, mix design, and bar type............................................................................................................ 55

Figure 28. Relationship between experimental and modeled corrosion rate of top bar in concrete ....................................................................................................... 56

Figure 29. Predicted top bar icorr as a function of deicer type, test type and bar type in the (a) bridge mix; (b) pavement mix ......................................................... 57

Figure 30. Predicted response surface of icorr of epoxy-coated dowel bar in pavement mix as a function of: (a) daily average Cl- concentration and wet percent time; (b) exposure duration and daily average exposure temperature ............ 60

Figure 31. Predicted response surface of icorr of stainless steel tube in pavement mix as a function of: (a) daily average Cl- concentration and wet percent time; (b) exposure duration and daily average exposure temperature ..................... 61

Figure 32. Predicted response surface of icorr of uncoated rebar in bridge mix as a function of: (a) daily average Cl- concentration and wet percent time; (b) exposure duration and daily average exposure temperature ........................... 62

Figure 33. Predicted response surface of icorr of uncoated rebar in bridge mix with embedded steel strip, as a function of: (a) daily average Cl- concentration and wet percent time; (b) exposure duration and daily average exposure temperature ..................................................................................................... 63

Figure 34. Ca elemental concentration of PCC samples following the Natural Diffusion Test with various solutions, with box plots illustrating the EDX data collected from multiple locations on the sample at two depths. ............. 69

Figure 35. Chemical composition of PCC samples following the Natural Diffusion Test with various solutions, with box plots illustrating the EDX data

ix

collected from multiple locations on the sample at two depths. (a) Si/Ca; (b) Al/Ca ......................................................................................................... 70

Figure 36. Chemical composition of PCC samples following the Natural Diffusion Test with various solutions, with box plots illustrating the EDX data collected from multiple locations on the sample at two depths. (a) S/Al; (b) S/Ca. .......................................................................................................... 71

Figure 37. Representative SEM image of PCC sample following the Natural Diffusion Test with non-inhibited NaCl. (a) pavement mix, 1 mm; (b) pavement mix, 12 mm; (c) and (d): both bridge mix, 12 mm, with magnification of 3510 times. .......................................................................... 72

Figure 38. Representative EDX spectrum and relevant elemental ratios of a pavement concrete sample following the Natural Diffusion Test with non-inhibited NaCl. Corresponding to area shown in Figure 36a, with magnification of 500 times. ........................................................................................................ 73

Figure 39. SEM image of PCC sample following the Natural Diffusion Test with inhibited NaCl. (a) pavement mix, 1 mm; (b) pavement mix, 12 mm; (c) and (d): both bridge mix, 12 mm, with magnification of 3510 times. ............ 74

Figure 40. Representative EDX spectrum and relevant elemental ratios of a pavement concrete sample following the Natural Diffusion Test with the inhibited NaCl. Corresponding to area shown in Figure 38a, with magnification of 500 times. ........................................................................................................ 75

Figure 41. SEM image of PCC sample following the Natural Diffusion Test with inhibited CaCl2. (a) pavement mix, 1 mm; (b) pavement mix, 12 mm; (c) and (d): both bridge mix, 12 mm, all with magnification of 3510 times except (a) at 500 times. ................................................................................... 76

Figure 42. A pavement concrete sample following the Natural Diffusion Test with the inhibited CaCl2, with magnification of 500 times: (a) SEM image; (b) corresponding EDX spectrum and relevant elemental ratios of the area. ....... 77

Figure 43. SEM image of PCC sample following the Natural Diffusion Test with inhibited MgCl2. (a) pavement mix, 1 mm; (b) pavement mix, 12 mm; (c) and (d): both bridge mix, 12 mm, all with magnification of 3510 times. ....... 79

Figure 44. Representative EDX spectrum and relevant elemental ratios of a pavement concrete sample following the Natural Diffusion Test with the inhibited MgCl2. Corresponding to area shown in Figure 42a, with magnification of 500 times. ........................................................................................................ 80

x

List of Tables

Table 1. Comparison of eutectic and effective temperatures for common deicers [12] ..... 7

Table 2. Information related to the actual liquids used for deicer ponding experiments .. 13

Table 3. Design of deicer ponding experiments ............................................................... 16

Table 4. Mix design used for this study: (a) pavement samples; (b) bridge samples. ...... 17

Table 5. Parameters and performance of the ANN models in this study .......................... 39

Table 6. Index values used to quantify the various influential factors ............................. 40

xi

Abbreviations and Acronyms

AASHTO American Association of State Highway and Transportation Officials ACR alkali-carbonate reaction AFm aluminate monosulfate hydrates AFt aluminate trisulfate hydrates

AgCl silver chloride

ANN artificial neural network ASR alkali-silica reaction ASTM American Society of Testing and Materials BP back-propagation C degrees Celsius C3A tricalcium aluminate C4AF Friedels salt, 3CaOAl2O3CaCl210H2O

Ca(OH)2 calcium hydroxide CaCl2 calcium chloride CMA calcium magnesium acetate CP cathodic protection C-S-H calcium silicate hydrate Dapp apparent diffusion coefficient Deff effective diffusion coefficient DI de-ionized water DOT Department of Transportation ECE electrochemical chloride extraction

Ecorr corrosion potential EDX energy dispersive x-ray spectroscopy EICI electrical injection of corrosion inhibitors EIS electrochemical impedance spectroscopy FDM Finite Difference Method FEM Finite Element Method F degrees Fahrenheit FESEM field emission scanning electron microscopy FHWA Federal Highway Administration FN Friction Number HPC high performance concrete

icorr corrosion current density

ITS indirect tensile strength ITZ interfacial transition zone KAc potassium acetate

KCl potassium chloride

xii

KFm potassium formate

L/D length/diameter ratio

LP linear polarization

LPR linear polarization resistance

MCI migrating corrosion inhibitor

MFP monofluorophosphate

Mg(OH)2 magnesium hydroxide

MgCl2 magnesium chloride MPY milli-inches per year M-S-H magnesium silicate hydrate NaAc sodium acetate NACE National Association of Corrosion Engineers NaCl sodium chloride NaFm sodium formate NCHRP National Center for Highway Research Program RCPT rapid chloride permeability test RMT rapid migration test SHRP Strategic Highway Research Program SMSE sum of mean square error OCP open circuit potential OGBM open-graded base material OGFM open-graded friction course OGM open-graded mix OGP open-graded pavement OH- hydroxyl OPC ordinary Portland Cement PAH polycyclic aromatic hydrocarbon PCC Portland cement concrete PCR percent corrosion rate PNS Pacific Northwest Snowfighters SAE Society of Automotive Engineers SCC self-compacting concrete SCE saturated calomel electrode SEM scanning electron microscopy SHRP Strategic Highway Research Program SMA stone matrix asphalt w/c water-to-cement ratio WTI Western Transportation Institute XPS X-ray photoelectron spectroscopy

1

E X E C UT I V E SUM M A R Y

Prior to this research, little was known about the possible contribution or risk corrosion inhibitors (and other additives) in deicers pose to infrastructure preservation. Therefore, research was needed to determine whether or not corrosion-inhibited deicers provide benefits in mitigating the corrosion of rebars or dowel bars in concrete, relative to the straight salt. To this end, we conducted an extensive literature review and performed laboratory tests to assess the effect of chloride-based deicers on rebars and dowel bars in concrete and to determine whether or not deicer corrosion inhibitors help preserve the transportation infrastructure.

The laboratory investigation exposed concrete samples to four common chloride-based deicers (assuming an average dilution factor of 100:31 from their eutectic concentration once applied) for approximately one year or less. The two mix designs investigated represent the roadway pavements and bridge decks built by the Washington State Department of Transportation before the 1980s, with a water-to-cement ratio of 0.39 and 0.38 and maximum aggregate size of 1.5 and 0.75 respectively. The bridge mix specimens were made with 0.1% sodium chloride by weight of concrete to simulate salt contamination; or with a steel strip to simulate crack; or without either; but all with number 4 AASHTO M-31 uncoated steel rebar in them. The pavement mix specimens were made with a simulated sawed joint and a MMFXTM dowel bar; or an epoxy-coated AASHTO M-284 dowel bar; or a stainless steel tube with epoxy-coated dowel bar insert. We aimed to simulate the effect of deicers on reinforced concrete in an accelerated manner, by either ponding the concrete samples with deicer solutions at room temperature, or incorporating pressurized ingress, wet-dry cycling and temperature cycling into the test regime. All concrete samples had a pond 2.5 deep in its upper portion. The concrete samples used for the Cyclic Exposure Test had two bars in them and the distance between the top bar and the bottom bar was 1, whereas the samples used for the Natural Diffusion Test only had one bar in them. The dowel bars had the dimension of length 18 diameter 1.5 and had a concrete cover of 2 over the top bar, whereas the rebars had the dimension of length 18 diameter 0.5 and had a concrete cover of 1.5 over the top bar.

The chloride ingress over time was monitored using a custom-made chloride sensor embedded in each concrete sample. Also periodically measured were the OCP of the top rebar or dowel bar and the macro-cell current flowing between the top bar and the bottom bar in concrete. Once the chloride sensor detected the arrival of sufficient chlorides near the top bar and the OCP data indicated the possible initiation of top bar corrosion, the corrosion rate of the top bar was also periodically measured using non-destructive, electrochemical techniques. At the end of ponding experiments, core specimens were taken from each sample to test the compressive strength of concrete. For samples continuously subjected to deicers at room temperature, microscopic analyses were conducted to examine the possible chemical changes in the cement paste.

The key findings and conclusions are presented as follows.

2

1. The laboratory testing revealed mixed effects of deicers on the compressive strength of Portland cement concrete. When the concrete specimens were continuously exposed to deicers at room temperature, the corrosion inhibitors added in deicers showed little benefits to the compressive strength of concrete. However, when the concrete specimens underwent temperature and wet/dry cycles, the corrosion inhibitors in deicers demonstrated a beneficial role in preserving the concrete strength. The inhibited magnesium chloride deicer posed a greater risk to concrete strength, relative to the other three deicers investigated, likely due to the role of magnesium cations in de-calcification of cement hydrates.

2. From a modeling perspective, artificial neural networks (ANNs) were used to achieve better understanding of the complex cause-and-effect relations inherent in the deicer/concrete/bar systems and were successful in finding some meaningful, logical results from the noisy data associated with the deicer ponding experiments.

3. The chloride sensor data and ANN modeling suggest that corrosion inhibitor (and

possibly other additives in the case of the inhibited calcium or magnesium chloride deicers) did slow down the ingress of the chlorides into concrete, for both natural diffusion and cyclic exposure tests. This benefit of corrosion inhibitor in deicers held true for all concrete mixes and was especially significant for concrete specimens subjected to the Cyclic Exposure Test or for bridge mix specimens with premixed sodium chloride or embedded steel strip. Among the four deicers investigated, the chloride anion (Cl-) associated with non-inhibited sodium chloride penetrated into concrete at the highest rate whereas that of the inhibited magnesium or calcium chloride deicers penetrated at much lower rates. The different chloride penetration rates have implications on the service life of reinforced concrete exposed to different deicers, even though such implications are difficult to quantify.

4. The electrochemical potential data and ANN modeling suggest that corrosion

inhibitor (and possibly other additives in the case of the inhibited calcium or magnesium chloride deicers) led to more positive top bar potentials at 298 days, implying the beneficial role of inhibitors in reducing the active corrosion risk of bars in concrete.

5. The macro-cell current data and ANN modeling suggest that corrosion inhibitors (and

possibly other additives in the case of the inhibited calcium or magnesium chloride deicers) helped to reduce the corrosion rate of the top bars in concrete. This benefit is not significant for the dowel bars in the pavement mix since none of them were actively corroding in the studied duration. According to the ANN modeling, at 254 days the dowel bars show consistently lower macro-cell current values than the rebars, despite their larger exposed surface area in concrete. This is likely due to the higher corrosion resistance of the dowel bars than that of the uncoated rebar, and also to the fact that the concrete cover over the top dowel bar in the pavement mix was 2 vs. 1.5 for the top rebar in the br idge mix. There were small differences in the macro-cell current data of various dowel bars investigated, limited by the short duration of this study.

3

6. The corrosion rate data of the top bar in concrete and ANN modeling suggest that in most cases, corrosion inhibitors (and possibly other additives in the case of the inhibited calcium or magnesium chloride deicers) reduced the corrosion rate of bars in concrete. The ANN modeling shows active corrosion of top rebar at 254 days in the bridge mix exposed to most deicers, whereas all top dowel bars seem to remain passivated (no active corrosion) in the pavement mix. The predicted corrosion rate of the top bar in concrete generally was the highest in non-inhibited sodium chloride and lowest in the inhibited magnesium chloride deicer. The benefit of deicer inhibitors for rebar in concrete was not significant for the bridge mixes with premixed sodium chloride or embedded steel strip where the rebar is actively corroding.

7. While the corrosion inhibitors in deicer products provide some benefits in delaying

the chloride ingress and subsequent corrosion initiation of steel in concrete, such benefits seem to diminish once the active corrosion of the rebar is initiated. In other words, the inhibitors showed little benefits in re-passivating the actively corroding rebars in concrete or in stifling the corrosion propagation.

8. We examined the effect of deicer ponding on the chemistry of concrete on the

microscopic level. The results revealed that each investigated deicer, at its given concentration, chemically reacted with some of the cement hydrates and formed new products in the concrete matrix. Such chemical changes of the cement paste induced by the deicers may account for the observed strength changes and have further implications for the concrete durability. Corrosion inhibitors and other additives in deicers did not show significant benefit in inhibiting the chemical changes of concrete induced by cations and/or anions in deicers.

Implementation Recommendations

1. Explore improved technologies and products and implement best practices to minimize the deicer usage while maintaining the desired levels of service.

2. Be aware of the deleterious effects various deicers could pose on the strength of concrete, especially in the case of magnesium-chloride-based deicers since they can chemically de-calcify the cement hydrates and undermine the concrete microstructure.

3. Be aware of the corrosion risk various deicers could pose on the rebar and dowel bar in concrete, especially in the case of sodium-chloride-based deicers (with or without inhibitor) since the chloride anions in them tend to ingress at higher speed than those in other deicers and thus induce rebar corrosion in concrete at an earlier stage.

4. Explore new technologies and methods to minimize the negative side effects of sodium chloride, magnesium chloride, and other deicers on reinforced concrete. To anticipate better return on investment in preserving the durability and performance of reinforced concrete, agencies should focus on improved concrete mix designs with

4

less permeability (e.g., replacing Portland cement with fly ash, silica fume, and slag), instead of procuring the more costly (inhibited) deicers.

5. Encourage the inclusion of a test method for deicer impact on reinforced concrete into the existing test matrix used to evaluate and qualify deicer products, to facilitate a more holistic approach to selecting snow and ice control chemicals through performance-based specifications.

5

C H A PT E R 1. I NT R ODUC T I ON

1.1. Problem Statement

For northern United States, Canada and other cold-climate regions, snow and ice control operations are essential to maintain roadway safety, mobility and productivity by providing safe driving surfaces in the winter season. Depending on the severity of winter, the Washington State Department of Transportation (WSDOT) typically spends $25-30 million dollars for the winter maintenance operations, and millions of pounds of deicers1

It is a popular practice to add corrosion inhibitors and other additives to deicer products, and the WSDOT Salt Pilot project has field evaluated the relative corrosivity of non-inhibited and inhibited chloride deicers to bare metals [1]. However, little is known about the possible contribution or risk deicer inhibitors (and other additives) pose to infrastructure preservation. Chloride ingress into concrete is a complex process, which in the highway environment is further complicated by the freeze-thaw cycles and wet-dry cycles experienced by roadways and bridges. Therefore, research is needed to determine whether or not corrosion-inhibited deicers provide benefits in mitigating the corrosion of rebars or dowel bars in concrete, relative to the straight salt (non-inhibited sodium chloride).

(mainly sodium chloride and corrosion-inhibited sodium/magnesium/calcium chlorides) are applied on the roadways every winter. Like in many northern states, there are growing concerns in the state of Washington over the corrosive effects that deicers pose on its multi-billion-dollar transportation infrastructure, including thousands of miles of roadways (especially concrete pavements with dowel bars) and thousands of bridges exposed to deicer applications.

1.2. Background

1.2.1. Chemicals for Winter Road Maintenance

Deicers applied onto highways often contain chlorides as freezing-point depressants because of their cost-effectiveness, including mainly sodium chloride (NaCl), magnesium chloride (MgCl2), and calcium chloride (CaCl2), sometimes blended with proprietary corrosion inhibitors. A recent survey of highway maintenance agencies conducted by the Western Transportation Institute (WTI) indicated that NaCl was the most frequently used deicer, followed by abrasives, then MgCl2, agriculture-based products, CaCl2, and others. Less than 25% of the survey respondents used alternative deicers such as potassium acetate (KAc), sodium acetate (NaAc), calcium magnesium acetate (CMA), and potassium formate [2].

NaCl is the most widely used chemical due to its abundance and low cost, whereas MgCl2 brines are often used in place of NaCl because laboratory tests have demonstrated their better ice-melting performance at lower temperatures [3]. Field studies have shown

1 For simplicity, the term deicer will refer to all chemicals for anti-icing, de-icing, and pre-wetting operations.

6

CaCl2 to be more effective than NaCl, owing to its ability to attract moisture and stay on the roads [4]. However, some agencies choose not to use CaCl2 as it does not dry and can cause roads to become slippery under certain conditions [5]. Chlorides are generally considered the most corrosive winter maintenance chemicals. Acetates such as KAc and CMA have also been used for highway anti-icing applications. While generally much more expensive, KAc and CMA can be more effective, less corrosive to carbon steel, and pose less environmental risks than chlorides [2]. Also available are a variety of agriculture-based chemicals used either alone or as additives for other winter maintenance chemicals [6]. Agriculture-based additives increase cost but may provide enhanced ice-melting capacity, and/or last longer than standard chemicals when applied on roads [7]. However, while some studies indicate that such products reduce corrosion [8], our recent study of a specific product (which had significant amount of chlorides in it) showed that it can be as corrosive as traditional deicers [2].

Maintenance agencies are continually challenged to provide a high level of service (LOS) and improve safety and mobility of winter roads in a cost-effective manner while minimizing corrosion and other adverse effects to the environment. In 2007 the U.S. sold approximately 20.2 million tons of deicing salts for use in winter maintenance [9]. The growing use of deicers has raised concerns over their effects on motor vehicles, transportation infrastructure, and the environment [2]. Over five billion dollars are spent each year by state and local agencies to repair infrastructure damage caused by snow and ice control operations. When using chloride-based deicers for road maintenance, the average cost due to corrosion and environmental effects are estimated to be at least three times as high as the nominal cost [10]. A recent National Cooperative Highway Research Program (NCHRP) report identified the deicer corrosion to steel rebar as the primary concern, followed by detrimental effects to vehicles, concrete in general, structural steel, and roadside structures [11].

Dowel bars are used to provide a mechanism to transfer wheel loads from one concrete pavement slab to the next in a reliable and cost-effective manner, in order to reduce joint deflection and minimize distresses such as faulting, pumping and cracking. Chloride-based deicers also pose a significant risk for the durability, serviceability and performance of concrete pavements with dowel bars embedded.

When selecting chemicals for winter road maintenance, agencies tend to consider their performance characteristics (e.g., effective temperature and ice melting capacity) along with their cost, application rates required for various road weather scenarios, and environmental risks (including those to metals and concrete). For deicers, one widely used tool to aid such decisions is its eutectic curve, which presents the eutectic temperature as a function of deicer concentration on the weight basis. Eutectic temperature is the minimum temperature that a deicer solution remains in liquid form. During the process of melting snow or ice, additional water is produced and the concentration of the deicer is reduced, which may cause the solution to re-freeze. Thus, the eutectic temperature can be significantly different from the effective temperature for a deicer (as shown in Table 1). In current practice, the effective temperature of deicers is generally determined by a consensus of field experience instead of a laboratory test.

7

Table 1. Comparison of eutectic and effective temperatures for common deicers [12]

Deicer

Eutectic Temperature, Teut

Minimum Effective Temperature, Teff

Source C F C F CaCl2 -51.6 -60.9 -35.0 -31.0 1 CaCl2 -51.1 -60 -28.9 -20 2 MgCl2 -33.3 -27.9 -20.0 -4.0 1 MgCl2 -33.3 -28 -15.0 5 2 Urea -11.7 10.9 -9.0 15.8 1 Urea -12.2 10 -3.9 25 2

Formamide -45.0 -49.0 -18.0 -0.4 1 NaCl -21.1 -6 -9.4 15 2 KAc -60.0 -76 -26.1 -15 2

CMA -27.2 -17 -6.1 21 2 CMA -10 14.0 -10 14.0 1

C2H6O2 (ethylene glycol) -51 -59.8 -23.3 -9.9 1

1 Resource Concepts Inc (1992) 2 Anonymous (2003)

1.2.2. Corrosion of Steel in Concrete and Concrete Permeability

For highway bridges and other reinforced concrete structures, the corrosion of reinforcing steel in concrete has been a major problem with serious economic and safety implications. Corrosive agents, liquid or gaseous, may penetrate the concrete through capillary absorption, hydrostatic pressure, or diffusion. The ingress of gases, water or ions in aqueous solutions into concrete takes place through pore spaces in the cement paste matrix and paste-aggregate interfaces or microcracks. For the durability of concrete, permeability is believed to be the most important characteristic [13], which is related to its microstructural properties, such as the size, distribution, and interconnection of pores and microcracks [14].

For reinforced concrete structures exposed to salt-laden environments, the chloride-induced corrosion of reinforcing steel is the major cause for their premature deterioration and degradation. Chloride ingress is one of the major forms of environmental attack for reinforced concrete bridges [15], which may induce corrosion of the reinforcing steel and a subsequent reduction in the strength, serviceability, and aesthetics of the structure. Chloride-induced corrosion of dowel bars is also a serious concern for reinforced pavements. Therefore, the chloride permeability of concrete has been recognized as a critical intrinsic property of the concrete [16].

There is a general agreement that the most effective measure for durability can be achieved at the design and materials selection stages by using adequate concrete cover and high quality concrete. Usually, an increase in concrete cover thickness leads to improved durability of reinforced concrete, as it enhances the barrier to the various aggressive species moving towards the reinforcement and increases the time for corrosion to initiate. In reality, however, the cover thickness cannot exceed certain limits, for mechanical and practical reasons [17].

8

1.2.3. Mechanisms of Steel Corrosion in Concrete

Concrete normally provides both chemical and physical protection for the steel reinforcement embedded in concrete. The cement hydration leads to the highly alkaline (pH 13-14) pore solution of concrete, which promotes the formation of an oxide/hydroxide film at the steel surface, a passive film of about 10 nanometers thick [18]. This protective film effectively insulates the steel and electrolytes so that the corrosion rate is negligible. In addition, the concrete cover prevents or at least retards the ingress of aggressive substances.

For the transportation infrastructure exposed to salt-laden environments, chloride ingress into concrete is of primary concern in terms of concrete durability. While the chloride ion (Cl-) has only a small influence on pore water pH, concentrations as low as 0.6 kilograms per cubic meter (kg/m3) (concrete weight basis) have been projected to compromise steel passivity [18]. Furthermore, the protection of steel by concrete is compromised by the gradual ingress of atmospheric carbon dioxide into the concrete, a process known as carbonation, which reduces the pore solution pH of carbonated concrete to the range of 8 to 9. The corrosion due to carbonation progresses at a much lower rate that that due to chloride ingress [19]. Concrete exposure to acids, sulfates and freeze-thaw cycles may also compromise the protection of steel.

Corrosion-induced concrete deterioration in the presence of Cl- generally proceeds in the following steps:

1. Ingress of chloride into the concrete, to a point where a critical concentration of Cl- results at the embedded reinforcement depth, known as the threshold chloride concentration, CT;

2. Local disruption of the passive film and onset of active corrosion, in the form of corrosion cells between the active corrosion zones (anode) and the surrounding still passive areas (cathode) [18,20];

Anode: Fe Fe+2 + 2 e- (or: Fe + 2Cl- FeCl2 + 2 e-)

Fe+2 + 2H2O Fe(OH)2 + 2 H+

Cathode: O2 + 2H2O + 4 e- 4OH-

Other possible reactions: Fe+2 + 2OH- Fe(OH)2

4Fe(OH)2 + O2 4-FeOOH + 2H2O

4Fe(OH)2 + 2H2O + O2 4Fe(OH)3

2Fe(OH)3 Fe2O3 + 3H2O

9

3. Accumulation of solid corrosion products (oxides/hydroxides) in the concrete pore space near the reinforcement and buildup of tensile hoop stresses around steel; and

4. Cracking or spalling of the concrete cover above the reinforcement.

Because of step 4, moisture, oxygen, and chlorides will gain more direct access to the embedded steel, and corrosion rate is further accelerated [21]. In step 2, for stable pit growth to be sustained, the relative concentrations of aggressive Cl- and inhibitive OH- should be above a certain ratio, otherwise repassivation will occur [22].

1.2.4. Transport of Chlorides and Inhibitors in Concrete: State of the Knowledge

This work started with a comprehensive literature review on topics relevant to this study. As detailed in Appendix A, we synthesized the information on existing research related to the transport of chlorides and inhibitors in concrete, which is valuable for understanding the chloride-induced corrosion of rebar or dowel bar in concrete. The various sections cover the measurement of chloride ingress into concrete, the challenges in assessing concrete durability from its chloride diffusivity, the inhibitor penetration into concrete, the computational models to simulate the transport of species in aqueous solution, water-unsaturated and water-saturated cementitious materials, and the service life modeling of reinforced concrete in salt-laden environments.

1.2.5. Deicer Impact on Concrete Performance and Durability

Both chloride-based and non-chloride-based deicers may pose detrimental effects on concrete infrastructure and thus reduce concrete integrity (as indicated by expansion, mass change and loss in the dynamic modulus of elasticity) and strength. Such risks of deicers on the durability of Portland Cement Concrete (PCC) structures and pavements exist through three main pathways: 1) physical deterioration of the concrete through such effects as salt scaling; 2) chemical reactions between deicers and cement paste (a cation-oriented process, especially in the presence of Mg2+ and Ca2+); and 3) deicers aggravating aggregate-cement reactions (including an anion-oriented process in the case of chlorides, acetates, and formates affecting alkali-silica reaction - ASR; and a cation-oriented process in the case of CaCl2 and MgCl2 affecting alkali-carbonate reaction - ACR). The use of proper air entrainment, high-quality cementitious materials and aggregates, and mineral admixtures is promising in mitigating the deicer impact on PCC. More details are provided in Appendix B and reference [23].

1.2.6. Deicer Corrosion to Metals in Transportation Infrastructure

Chemical deicers, especially those based on chlorides, may cause corrosion damage to the transportation infrastructure such as reinforced or pre-stressed concrete structures and steel bridges. One study has estimated that the use of road salts impose infrastructure corrosion costs of at least $615 per ton, vehicular corrosion costs of at least $113 per ton, aesthetic costs of $75 per ton if applied near environmentally sensitive areas, in addition to uncertain human health costs [24]. The estimated cost of installing corrosion protection measures in new bridges and repairing old bridges in the Snowbelt states is between $250

10

million and $650 million annually [25]. Parking garages, pavements, roadside hardware, and non-highway objects near winter maintenance activities are also exposed to the corrosive effects of road salts. It should be noted that any repairs to the infrastructure translate to costs for the user in terms of construction costs, traffic delays and lost productivity. Indirect costs are estimated to be greater than ten times the cost of corrosion maintenance, repair and rehabilitation [26].

For practical purposes, all chloride-based deicers were ranked equally high in terms of corrosion risk to the reinforcing steel in a recent NCHRP report [11], even though hygroscopic chlorides of magnesium and calcium can be more aggressive to the exposed metals than NaCl because of the longer time of wetness. The cation (Na+, Ca2+, or Mg2+) associated with Cl- also affects the pH value of the electrolyte and the chloride diffusion coefficient in concrete poses different levels of corrosion risk to the rebar in concrete. The relative corrosivity of deicers is dependent on many details related to the metal/deicer system. Therefore, no general conclusions should be made when ranking corrosion risks of different deicer products. Instead, it is important to note the test protocol employed, the metal coupons tested, the deicer concentrations, the test environment, etc. It is also extremely difficult to relate laboratory test results of corrosion resistance to the actual field performance of metals. There are many ways to manage the corrosive effects of deicers, such as: selection of high-quality concrete, adequate concrete cover and alternative reinforcement, control of the ingress and accumulation of deleterious species, injection of beneficial species into concrete, and use of non-corrosive deicer alternatives and optimal application rates. More details are provided in Appendix C and reference [27].

1.3. Study Objectives

The objectives of this research are to assess the effect of chloride-based deicers on rebars and dowel bars in concrete and to determine whether or not deicer corrosion inhibitors help preserve the transportation infrastructure. To this end, this research includes a comprehensive literature review on relevant topics and a laboratory investigation using reinforced concrete samples exposed to deicers (sodium chloride and corrosion-inhibited sodium/magnesium/calcium chlorides, assuming an average dilution factor of 100:31 from their eutectic concentration once applied). The laboratory investigation aims to simulate the effect of deicers on reinforced concrete in an accelerated manner, by either ponding the concrete samples with deicer solutions at room temperature, or incorporating pressurized ingress, wet-dry cycling and temperature cycling into the test regime. The chloride ingress over time will be monitored using a custom-made chloride sensor embedded in each concrete sample. Also periodically measured are the open circuit potential (OCP) of the top rebar or dowel bar and the macro-cell current flowing between the top bar and the bottom bar in concrete. Once the chloride sensor detects the arrival of sufficient chlorides near the top bar and the OCP data indicate the possible initiation of top bar corrosion, the corrosion rate of the top bar will also be periodically measured using non-destructive, electrochemical techniques. At the end of ponding experiments, core specimens will be taken from each sample to test the compressive strength of concrete. For samples continuously subjected to deicers at room temperature,

11

microscopic analyses will be conducted to examine the possible chemical changes in the cement paste.

1.4. How This Report Is Organized

The following chapter will discuss the methodology used in gathering and analyzing data from the laboratory investigation. Chapter 3 presents the chloride diffusion coefficient of various deicers in concrete, the compressive strength data of deicer-contaminated concrete and the laboratory testing results from deicer ponding experiments and associated analyses based on predictive models. Chapter 4 presents the Scanning Electron Microscopy/Energy-dispersive X-ray Spectroscopy (SEM/EDX) data for the concrete samples subjected to natural diffusion of deicers and the subsequent analysis of chemical changes of the cement paste. In general, each chapter is dedicated to a single theme. Finally, Chapter 5 summarizes the key findings from the previous five chapters and presents research needs identified from this project, followed by suggestions and recommendations for implementation by WSDOT. Appendices conclude this report.

1.5. References

[1] E.V. Baroga, 2002-2004 Salt Pilot Project. Final Report for the Washington State Department of Transportation. 2005.

[2] Shi, X., Fay, L., Gallaway, C., Volkening, K., Peterson, M.M., Pan, T., Creighton, A., Lawlor, C., Mumma, S., Liu, Y., and Nguyen, T.A. Evaluation of Alternate Anti-icing and Deicing Compounds Using Sodium Chloride and Magnesium Chloride as Baseline Deicers. Final Report for the Colorado Department of Transportation. Denver, CO. Report No. CDOT-2009-01. Feb. 2009.

[3] Ketcham, S.A., L.D. Minsk, R.R. Blackburn, and E.J. Fleege (1996). Manual of Practice for an Effective Anti-Icing Program: A Guide for Highway Winter Maintenance Personnel. Publication No. FHWA-RD-9-202. Army Cold Regions Research and Engineering Laboratory.

[4] Warrington, P.D (1998). Roadsalt and Winter Maintenance for British Columbia Municipalities. Best Management Practices to Protect Water Quality. Environmental Protection Agency, December 1998.

[5] Perchanok, M.S., D.G. Manning and J.J. Armstrong (1991). Highway De-Icers: Standards, Practices, and Research in the Province of Ontario. Research and Development Branch MOT. Mat-91-13.

[6] Nixon, W.A. and A.D. Williams (2001). A Guide for Selecting Anti-icing Chemicals. Ver. 1.0. IIHR Technical Report No. 420.

[7] Fischel, M. (2001). Evaluation of Selected Deicers Based on a Review of the Literature. The SeaCrest Group. Louisville, CO. Report Number CDOT-DTD-R-2001-15.

[8] Kahl, S. (2004).Agricultural By-Products for Anti-Icing and De-Icing Use in Michigan. SNOW04-009. Sixth International Symposium on Snow Removal and Ice Control Technology. Transportation Research Circular E-C063: Snow and Ice Control Technology. June 2004. pp 552-555.

[9] Salt Institute. http://www.saltinstitute.org/Production-industry/Facts-figures/U.S.-salt-production-sales. Accessed in July 2010.

12

[10] Shi, X. The Use of Road Salts for Highway Winter Maintenance: An Asset Management Perspective. 2005 ITE District 6 Annual Meeting. Kalispell, Montana. July 10-13, 2005.

[11] Levelton Consultants Ltd, Guidelines for the selection of snow and ice control materials to mitigate environmental impacts, NCHRP REPORT 577, National Cooperative Highway Research Program, Transportation research board of the national academies, Washington, D.C.

[12] Akin, M., and Shi, X. Development of Standardized Test Procedures for Evaluating Deicing Chemicals. Final report prepared for the Wisconsin Department of Transportation and the Clear Roads Program. March 2010.

[13] Baykal, M., Implementation of Durability Models for Portland Cement Concrete into Performance-Based Specifications. Austin, TX: The University of Texas at Austin. 2000.

[14] Savas, B., Effects of Microstructure on Durability of Concrete. Raleigh, NC: North Carolina State University. 1999.

[15] Samples, L. and J. Ramirez, Methods of Corrosion Protection and Durability of Concrete Bridge Decks Reinforced with Epoxy-Coated Bars, Phase I. FHWA/IN/JTRP-98/15. Purdue University, IN. 1999.

[16] Wee, T., A. Suryavanshi, and S. Tin, ACI Materials Journal, 97(2000): 221. [17] Bertolini L., B. Elsener, P. Pedeferri, and R. Polder. Corrosion of Steel in Concrete:

Prevention, Diagnosis, Repair. Wiley-VCH, Verlag GmbH & Co. KgaA, Weinheim. 2004.

[18] Hartt, W. and J. Nam, Critical Parameters for Corrosion Induced Deterioration of Marine Bridge Substructures in Florida, prepared for the Florida Department of Transportation. 2004.

[19] Basheer L., J. Kropp, and D.J. Cleland. Construction and Building Materials 15(2001): 93.

[20] Hausmann, D., Materials Performance 37(1998): 64. [21] Hartt, W., S. Charvin, and S. Lee, Influence of Permeability Reducing and

Corrosion Inhibiting Admixtures in Concrete upon Initiation of Salt Induced Embedded Metal Corrosion, prepared for the Florida Department of Transportation. 1999.

[22] Page C.L., N.R. Short, and W.R. Holden, Cement and Concrete Research 16(1986): 79.

[23] Shi, X., Akin, M., Pan. T., Fay, L., Liu, Y., and Yang, Z., The Open Civil Engineering Journal, 3(2009): 16-27.

[24] D. Vitaliano, J. Policy Anal. Mgmt. 11(3) (1992), 397-418. [25] Transportation Research Board, Highway de-icing: comparing salt and calcium

magnesium acetate. National Research Council. Special Report 235, 1991. [26] M. Yunovich, N.G. Thompson, T. Balvanyos, and L. Lave, Corrosion costs of

highway bridges, 2002. http://www.corrosioncost.com/pdf/highway.pdf, accessed in July 2007.

[27] Shi, X., Fay, L., Yang, Z., Nguyen, T.A., and Liu, Y. Corrosion Reviews, 27(1-2)(2009): 23-52.

13

C H A PT E R 2. M E T H ODOL OG Y

2.1. Experimental

2.1.1. Deicers of Interest

This study involved four liquid deicers of interest to WSDOT. The non-inhibited NaCl brine served as a control and was a 23% aqueous solution prepared by the research team, using solid rock salt provided by the Montana Department of Transportation (MDT), Bozeman District, from the salt stockpile. The corrosion-inhibited NaCl brine was prepared by adding a given amount of Shield GLTTM inhibitor into the non-inhibited NaCl brine and stirring to blend (as specified by the vendor). The corrosion-inhibited MgCl2 brine, Freezgard CI PlusTM was provided by the MDT-Bozeman district from their liquid storage tanks and contained approximately 30% MgCl2 by weight of the solution. The corrosion-inhibited CaCl2 brine, Geomelt CTTM and was provided by WSDOT directly through America West Environmental Supplies, Inc., and contained approximately 30% CaCl2 by weight of the solution along with some other chlorides (MgCl2, NaCl and KCl) as integral part of the formulation. On average, these liquid deicers were further diluted by a factor of 100:31 before being used for the deicing ponding experiments. The average daily concentrations of the deicer solutions used for the deicing ponding experiments are provided in Table 2. Note that the average daily salt and Cl- concentrations for the Cyclic Exposure Test was lower than those for the Natural Diffusion Test, since the former incorporated dry cycles featuring zero concentrations in the pond.

Table 2. Information related to the actual liquids used for deicer ponding experiments

Average Daily Salt

Concentration (by weight of

solution)

Dilution Ratio from Eutectic Concentration

Average Daily Cl-

Concentration (M)

Average Daily Salt

Concentration (by weight of

solution)

Dilution Ratio from Eutectic Concentration

Average Daily Cl-

Concentration (M)

Non-inhibited NaCl

NaCl 58.44 23% 7.7% 33% 1.31 6.38% 28% 1.09

Inhibited NaCl NaCl 58.44 23% 7.8% 34% 1.33 6.38% 28% 1.09

Inhibited CaCl2

CaCl2 110.98 30% 8.2% 27% 1.48 7.07% 24% 1.27

Inhibited MgCl2

MgCl2 95.21 30% 8.8% 29% 1.86 7.61% 25% 1.60

Average 8.1% 31% 1.49 6.9% 26% 1.26

Natural Diffusion Test Cyclic Exposure Test

Deicer

Main Freezing-

point Depressant

Molecular Weight (g/mol)

Eutectic Concentration (by weight of

solution)

The Pacific Northwest Snowfighters (PNS), an Association of transportation professionals for British Columbia, Washington, Idaho, Montana, Oregon, and Colorado, has implemented testing protocols and guidelines for new product qualification for deicers. A central feature of these requirements is the presence of corrosion inhibitor(s) in

14

deicers, and the qualification and evaluation of all deicers by a modified National Association of Corrosion Engineers (NACE) corrosion test.

As such, for this study we tested the deicer corrosivity in our laboratory before using them to conduct the ponding experiments of reinforced concrete samples. The corrosion tests followed the gravimetric method as specified by the PNS Association using the modified NACE Standard TM0169-95, aimed to assess the 72-hour average corrosivity of a deicer solution to carbon steel. The PNS Association has modified this procedure so that the test procedure uses 30 ml of a 3% chemical product solution as received per square inch of coupon surface area for the corrosion test. The PNS/NACE corrosion test entailed cyclic immersion (10 minutes in the solution followed by 50 minutes exposed to air) of multiple parallel steel coupons for 72 hours on a custom-design machine (by AD-TekTM), followed by a measurement of weight loss. The weight loss result in MPY (milli-inch per year) was translated into a percentage, or percent corrosion rate (PCR), in terms of the solution corrosivity relative to a eutectic reagent-grade NaCl brine. The coupons used were TSI washer steel and met ASTM F 436, Type 1, with a Rockwell Hardness of C 38-45. The test results shown in Figure 1 provide the baseline of the four deicers used in this study and indicate that the corrosion-inhibited deicers had much lower corrosivity than the non-inhibited salt brine (control). In other words, the inhibitors added in deicer products were effective in mitigating the corrosive attack of chloride to carbon steel under the investigated laboratory conditions. Note that the corrosivity data correspond to the deicer solutions diluted from their individual eutectic concentration by a factor of 100:3 (as specified by the PNS Association).

Relative Corrosivity of Deicers to Carbon Steel

94.0

18.5 19.9 17.0

0

20

40

60

80

100

NaCl I-NaCl I-CaCl2 I-MgCl2

Deicer Type

Perc

ent C

orro

sion

Rat

e (b

y th

e PN

S/N

AC

E M

etho

d)

F igur e 1. R elative cor r osivity of the four deicer s used in this study, based on the PNS/NA C E method.

15

2.1.2. Design of Deicing Ponding Experiments

To investigate the potential impact of deicers to rebar or dowel bar in concrete, this study included the following concrete/bar combinations of interest to WSDOT.

a. One sound concrete mix representing bridge decks with uncoated steel rebar inside;

b. One cracked concrete mix representing bridge decks with uncoated steel rebar inside;

c. Same as a except with 0.1% NaCl (by weight of concrete) admixed in fresh concrete;

d. One concrete mix representing pavement concrete with a simulated sawed joint and a MMFXTM dowel bar 2

e. One concrete mix representing pavement concrete with a simulated sawed joint and epoxy-coated AASHTO M-284 dowel bar;

;

f. Same mix as "d" except for a stainless steel tube with epoxy-coated dowel bar insert 3

Note that for the combination b, we did not use mechanical force to generate the cracks since such methods (e.g., compression) would generate cracks of non-reproducible nature. Instead, we vertically embedded a carbon steel strip (with thickness of 5 mm, length of 50 mm and width of 25 mm) in the concrete sample, with its upper side directly exposed to deicers and its lower side 25 mm into the concrete. As such, during the deicer ponding experiments, we anticipate the generation of micron-scale cracks at the steel strip/concrete interface through the corrosion of the steel strip and the subsequent formation of corrosion products.

.

Table 3 presents the overall experimental design used in this study. The reinforced concrete samples were exposed to deicers by either ponding at room temperature (referred to as Natural Diffusion Test in this report), or by incorporating pressurized ingress, wet-dry cycling and temperature cycling into the test regime (referred to as Cyclic Exposure Test in this report). As shown in Table 3, a total of 72 reinforced concrete samples were used in the Cyclic Exposure Test (three duplicates for each deicer/concrete/bar combination) and a total of 16 reinforced concrete samples were used in the Natural Diffusion Test (two duplicates for each deicer/concrete combination, using only uncoated rebar).

2 These were MMFX bars, instead of MMFX-2 bars. 3 The materials properties data of the bars are available from the WSDOT vendors. All the dowels used in this project met the WSDOTs standard specifications for such products.

16

Table 3. Design of deicer ponding experiments

Cyclic Exposure Test: 3 replicates each (totally 72 specimens in the environmental chamber)

Deicers

NaCl I-NaCl I-CaCl2 I-MgCl2

Bridge Mix

Uncoated Rebar 3 3 3 3 Uncoated Rebar w/

steel strip in concrete 3 3 3 3

Uncoated Rebar w/ 0.1wt.% admixed NaCl 3 3 3 3

Pavement Mix

Epoxy Coated Dowel 3 3 3 3 Stainless Steel Tube w/

Epoxy Coated Insert 3 3 3 3

MMFX Dowel Bar 3 3 3 3 Natural Diffusion Test: 2 replicates each (totally 16 specimens at room temperature)

Deicers

NaCl I-NaCl I-CaCl2 I-MgCl2 Bridge Mix Uncoated Rebar 2 2 2 2

Pavement Mix Uncoated Rebar 2 2 2 2

2.1.3. Concrete Samples: Materials and Preparation

For preparing the concrete samples used in this study, the mixing water, type I-II cement, coarse and fine aggregates, chemical admixtures, uncoated number 4 AASHTO M-31 rebars, epoxy-coated AASHTO M-284 dowel bars 4

The concrete samples used for the Cyclic Exposure Test had two bars in them and the distance between the top bar and the bottom bar was designed to be 1 (25.4 mm). The reinforced concrete samples used for the Natural Diffusion Test only had one bar in them. The thickness of concrete cover for both types of samples was determined in consultation with WSDOT, considering two main constraints. On the one hand, the cover should be thick enough to simulate actual bridge decks and pavements as well as accommodate the maximum aggregate size and the diameter of the bar in them. On the other hand, it is desirable to reduce the cover thickness in order to reduce the amount of time required for sufficient chlorides to reach the top bar surface and initiate its active corrosion.

, MMFX dowel bars, and stainless steel tubes with epoxy-coated dowel bar insert were obtained by WSDOT from various sources. The specific materials properties data of these bars are available from the individual WSDOT vendors per request.

4 Note that these dowel bars were of green color. The purple dowels shown in Figure 3a were ASTM A-934 epoxy dowels provided by a vendor for testing as add-on to this project.

17

This study used two concrete mixes, both obtained from Stoneway Concrete Inc., as shown in Table 4. The concrete mixes had no fly ash included, in order to simulate the WSDOT roadway pavements and bridge decks built before 1980. Note that the designed water-to-cement (w/c) ratio was 0.39 and 0.38 for the pavement and bridge mixes respectively.

Table 4. Mix design used for this study: (a) pavement samples; (b) bridge samples.

The dowel bars had the dimension of length 18 diameter 1.5 (LD: 457.238.1 mm) and had a concrete cover of 2 (50.8 mm) over the top bar 5

The optimum dimensions of the concrete samples were selected considering all the constraints imposed by the optimum design, such as the maximum aggregate size, weight of the cured specimen, bar dimensions and locations, transport distance of deicer solution from the pond to the top bar, and location of the chloride sensor. The pavement concrete samples used for the Cyclic Exposure Test had a rectangular shape and final dimensions of 16129 (LWH: 406305229 mm), with a pond 2.5 (63.5 mm) deep in its

. In light of the fact that most mix designs for WSDOT roadway pavements have a maximum aggregate size of 1.5 (38.1 mm) to guarantee sufficient strength, a pavement mix design with a maximum aggregate size of 1.5 (38.1 mm) was used. The rebars had the dimension of length 18 diameter 0.5 (LD: 457.212.7 mm) and had a concrete cover of 1.5 (38.1 mm) over the top bar. A bridge mix design with a maximum aggregate size of 0.75 (19 mm) was used to minimize the sample size and to minimize the risk of hydration cracking.

5 In this study, the fabrication of concrete specimens with dowel bars did not include a crack to simulate a transverse joint. In other words, the sawed joint was embedded in concrete. As such, the research is focused on the penetration of deicers through the concrete pavement before they reach the dowel bar surface. In practice, however, dowel bars could be in direct contact with the deicers applied in the environment, through the transverse joints.

18

upper portion. A custom-made wooden mold used for fabricating such samples is shown in Figure 2a.

(a)

(b)

(c)

F igur e 2. (a) A custom-made wooden mold used to fabr icate the r einfor ced pavement samples; (b) a por tion of the mold to make the pond; (c) a fabr icated

concr ete sample with the ponding mold not yet r emoved

This designed mold gave the fabricated concrete samples walls of 3 (76 mm) thickness around the pond, with reinforcing bolts in the walls. This design survived a pressure of 10 psi during the mock tests without showing any noticeable cracking and maintained the pressure in the enclosed pond relatively well. The pond was made by embedding a wooden mold (see Figure 2b) in the upper portion of the concrete specimen during casting, and then removing it two days after the concrete casting, leaving a pond with dimension of 1062.5 (254152.463.5 mm) in length, width and height respectively. Figure 3c shows a fabricated concrete sample with the ponding mold not yet removed. Note that a plastic ponding mold would have greatly facilitated its de-molding.

In addition, the bridge concrete samples used for the Cyclic Exposure Test had a rectangular shape and final dimensions of 16127.5 (LWH: 406305190 mm). The reinforced concrete samples used for the Natural Diffusion Test had a rectangular shape and final dimensions of 16127.5 (LWH: 406305190 mm) and 16125 (406305127 mm) respectively, for the pavement and bridge mixes respectively, with a pond 2.5 (63.5 mm) deep in its upper portion.

19

In October 2006, the concrete specimens were cast at a job site of Stoneway Concrete Inc. in Seattle, WA, by the research team with assistance from Stoneway personnel 6

Figure 3

. All concrete specimens, according to the design of experiments in Table 3, were fabricated with the received rebars, dowel bars, as well as the custom-made specimen molds and custom-made chloride sensors. shows some of the concrete samples during and after the fabrication process. To allow some redundancy, a total of 78 samples were cast for the pavement mix and 63 were cast for the bridge mix 7.

(a) 8

(b)

F igur e 3. Some pavement samples (a) dur ing and (b) after the fabr ication pr ocess (including pour ing and vibr ation).

6 The process of concrete casting was also witnessed by a WSDOT representative, Jim Weston from the WSDOT Material Laboratory, who also provided valuable advice for ensuring the quality of concrete samples. 7 These numbers do not account for the additional samples fabricated for the add-on project. 8 Note that the purple dowels shown here were ASTM A-934 epoxy dowels provided by a vendor for testing as add-on to this project.

20

Slump and air content measurements were performed to check the workability and quality of the freshly mixed concrete, as shown in Figure 4. For the pavement and bridge mixes, the slump was tested as 3.5 (89 mm) and 4.75 (121 mm) respectively, and both values fell in the specified ranges as provided by Stoneway (2.5 4.5 or 38 89 mm for pavement mix and 4.0 5.0 or 102 127 mm for bridge mix). The air content was tested as 5.7% and 5.1% for the pavement and bridge mixes respectively, and both values fell in the specified range (3.5% - 6.5%). Six cylinder specimens for compressive strength test were also prepared for quality control of each mix, which showed average 7-day strength of 3943 psi (27.2 MPa) and 4318 psi (29.8 MPa) for the pavement and bridge mixes respectively, indicating reasonable quality of these fabricated concrete samples. The deicer ponding experiment did not start until the concrete specimens were at least 60 days old to allow sufficient hydration.

(a)

(b)

(c)

F igur e 4. Quality contr ol of fabr icated concr ete samples: (a) slump test; (b) air content test; and (c) cylinder molds for fabr icating compr essive str ength test

specimens.

21

2.1.4. Chloride Sensors

The chloride ingress over time was monitored using a custom-made chloride sensor embedded in each concrete sample. Chloride sensors made of silver/silver chloride (Ag/AgCl) electrodes were fabricated in our laboratory using an electro-deposition process. First, the silver wire was sanded and cleaned, followed by the measurement of its diameter. Then, it is mounted in a 0.2M potassium chloride (KCl) solution along with a clean graphite rod, with them connected to the positive and negative end of a computer-controlled Princeton Applied ResearchTM Potentiostat (serving as the power source). A constant current is applied by the Potentiostat to provide a 1 mA/cm2 current density on the silver wire for one hour (to deposit a silver chloride layer on it), followed by the rinsing of the electroplated silver wire.

The fabricated Ag/AgCl electrodes were stored in saturated KCl solution in a dark glass bottle. As shown in Figure 5, each chloride sensor (dark color) was welded to a copper wire at each end (the welding points sealed with epoxy resin) and then carefully rinsed before being embedded above the top bar in the concrete samples. The chloride sensor was fixed at 5 mm above the top bar instead of at the same vertical height as the top bar, as a result of the specific configuration we chose and the effort to avoid any possible physical contact of the sensing layer with the top bar.

F igur e 5. C ustom-made chlor ide sensor s (dar k color ) welded to copper wir es (r ed color ), befor e being embedded in the concr ete samples.

The quality of these custom-made chloride sensors were first examined by reading their electrochemical potential in aqueous solutions containing known amounts of NaCl. As shown in Figure 6a, the chloride sensor calibration curve established in such neutral solutions showed a strong linear correlation between the natural logarithm of the chloride concentration and the chloride sensor potential (ECl-, vs. an Ag/AgCl/saturated KCl reference electrode). The high R-square value and the excellent reproducibility (between sensor 1 and sensor 2 in the figure) confirmed the quality of these chloride sensors.

The chloride sensor calibration curve was established in highly alkaline simulated concrete pore solutions (shown in Figure 6b), which differed from the calibration curve with neutral solutions, suggesting the joint contribution of hydroxyls (OH-) and Cl- to the electrochemical potential of the sensor.

22

Calibration of Custom-made Chloride Sensor in Netural Solutions

y = -21.968x + 36.831R2 = 0.9987

y = -21.617x + 37.254R2 = 0.9998

0102030405060708090

100

-2.7 -2.4 -2.1 -1.8 -1.5 -1.2 -0.9 -0.6 -0.3 0 0.3 0.6 0.9

Ln [Cl-], M

E Cl-,

mv

(vs.

Ag/

AgC

l)

Sensor 1Sensor 2Linear (Sensor 2)Linear (Sensor 1)

(a)

Calibration of Custom-made Chloride Sensor in Simulated Concrete Pore Solutions

y = -16.299x + 25.411R2 = 0.9992

0

30

60

90

120

150

-8 -6 -4 -2 0

Ln [Cl-], M

E Cl-,

mv

(vs.

Ag/

AgC

l) Sensor 1Linear (Sensor 1)

(b)

F igur e 6. E lectr ochemical potential of custom-made chlor ide sensor s as a function of chlor ide concentr ation in (a) neutr al solutions; (b) highly-alkaline simulated

concr ete por e solutions.

Finally, electrochemical potential of such sensors was tested by embedding them in concrete samples featuring a mix design similar to Table 4a but with known amounts of NaCl admixed. Figure 7 presents the ECl- reading at 28 days of concrete curing as a function of the concentration of the admixed NaCl by weight of cement, which shows relatively high variability (indicated by the low R-square value) and further deviation from Figure 6a. This is consistent with existing research, suggesting the oxidation of the sensing Ag/AgCl layer by the highly alkaline environment. While not ideal, the Ag/AgCl electrodes still served as the chloride sensor embedded in concrete.

23

Calibration of Custom-made Chloride Sensor in Embedded in Concrete

y = -41.86x + 28.218R2 = 0.627

-100

-60

-20

20

60

100

140

180

-4 -2 0 2 4

Ln [NaCl%], by weight of cement

E Cl-,

mv

(vs.

Ag/

AgC

l) Sensors at time of 28 days

Linear (Sensors at time of 28 days)

F igur e 7. E lectr ochemical potential of custom-made chlor ide sensor s as a function of NaC l concentr ation in 28-day concr ete samples.

2.1.5. Deicing Ponding Experiments

The Cyclic Exposure Test established and utilized a pressurized transport method, which aimed to accelerate the ingress of deicers into concrete and also to simulate the field scenario of traffic flow on pavements and bridge decks. This method forces the flow of chlorides and corrosion inhibitors (and other additives) into concrete by exposing one face of the concrete to the deicer solution that is under pressure. All other faces of the concrete are sealed. This served to drive the chlorides and inhibitors into the concrete under both convection and diffusion, governed by the following equation [1]:

(1)

where v is the average linear rate of flow which is:

(2)

and k is the hydraulic permeability, n is the porosity, and h is the applied pressure head. The solution to this differential equation is:

(3)

24

This allows the determination of diffusion coefficients for chlorides and inhibitors, when a concentration profile was known at a specific time.

To maintain a constant pressure in the deicer ponds, all specimens were connected to an air compressor via a manifold, vinyl tubing, and hose barbs. Figure 9 shows the schematics and experimental setup for the pressurized transport of deicers into concrete, with concrete samples attached to the pressure manifold within an environmental chamber at Montana State University Bozeman.

(a)

(b)

F igur e 8. (a) Schematics and (b) exper imental setup for the pr essur ized tr anspor t of deicer s into concr ete

25

In the experimental setup, air temperature was precisely regulated and closely monitored. In this study, the applied pressure was maintained at 3 psi during the wet cycles of the Cyclic Exposure Test. No pressure was applied during the dry cycles, during which we assumed that the externally applied pressure would make little difference in the transport of species in concrete. Additionally, the temperature cycles were defined and maintained using the environmental chamber. Figures 9a and 9b show the environmental chamber and the reinforced concrete samples for the Cyclic Exposure Test on the shelves inside the chamber. The temperature cycles were designed as follows: warm (52F) cold (20F) warm (52F) hot (84F) cycling on a weekly basis, intended to simulate the varied seasons in Washington in an accelerated manner. Along with varying temperatures, the specimens underwent wet-dry cycles. The deicers were added to each specimen for 3 weeks, and then removed for 1 week. Such cycles were repeated until termination of the specimen with little interruption, as shown in Figure 10. Note that the relative humidity inside the environmental chamber fluctuated as a function of air temperature, that is, lower temperature tended to cause lower relative humidity of the air.

(a)

(b)

F igur e 9. (a) The environmental chamber used for this study; (b) The reinforced concrete samples for the

Cyclic Exposure Test on the shelves inside the chamber.

In addition to the Cyclic Exposure Test, some reinforced concrete samples were subjected to the Natural Diffusion Test (as specified by Table 3). They were maintained at room temperature in the WTI laboratory facility, featuring air temperature of 733F and relative humidity of 505 percent.

26

Cycle vs Time

0

10

20

30

40

50

60

70

80

90

17 35 56 77 98 119 140 161 182 203 224 245 266 287 308 329 350 371

Days Since Average Start Date (7/02/07)

Cycl

e Te

mpe

ratu

re (o

F)

- D

ry -

Wet

F igur e 10. T emper atur e and wet/dr y cycles exper ienced by the r einfor ced concr ete samples dur ing the C yclic E xposur e T est

During the deicer ponding experiments, the electrochemical potential of the chloride sensor was periodically measured, using an Ag/AgCl/saturated KCl reference electrode in a wet sponge9

Similarly, the OCP of the top bar in concrete were monitored to provide additional indication of chloride arrival and possible initiation of active corrosion (also known as pitting corrosion) at the top bar. According to ASTM C876 guidelines, the probability of corrosion initiation is greater than 90% when the OCP of steel in concrete is more negative than 270 mV, relative to the saturated calomel electrode (SCE), which equals -235 mV relative to the Ag/AgCl/saturated KCl electrode. This rule, however, may not be applicable for water-saturated concrete or dry concrete, where the availability of dissolved oxygen or moisture can significantly affect the measured OCP of steel in

placed on the concrete exterior surface adjacent to the sensor. The chloride sensor was used to detect the arrival of chlorides near the top bar, since its potential (ECl-) indicates the chloride ion activity at that specific concrete depth. Generally lower ECl- values correspond to higher concentrations of free chloride ions, but such an assumed relationship might not hold when dealing with a malfunctioning sensor due to its oxidation in concrete or when the resistance between the reference electrode (placed in wet sponge) and the chloride sensor was large.

9 To allow good electric contact for the monitoring of chloride sensor and top bar, the epoxy coating for a small area at the side of each concrete sample was removed in order to place the wet sponge.

27

concrete. The threshold value will also change as a function of the temperature of service environment. Generally lower OCP values of the top bar correspond to higher risk of active corrosion, but such an assumed relationship might not hold when dealing with epoxy-coated bars or when the resistance between the reference electrode (placed in wet sponge) and the top bar was large.

For concrete samples subjected to the Cyclic Exposure Test, macro-cell current was also monitored to confirm whether any passive-to-active potential shift was in fact a consequence of top bar corrosion. As shown in Figure 8a, two rebars or dowel bars of the same type were embedded in the concrete samples, using a 1 (25.4 mm) vertical spacing. Periodically, the current between the pair of bars was monitored with a zero resistance ampermeter. When the chloride concentration reached the threshold level at the top bar depth and initiated active corrosion, the top bar would feature a lower OCP than the bottom bar. As such, the dynamics of macro-cell current was used to detect the onset of active corrosion of the steel, and provided an indirect measure of corrosion rate, similar to previous research [2] as shown in Figure 11. For this study, a macro-cell current of 10 A and 40 A was used as the threshold for closer monitoring of the rebar and dowel bars respectively. This translated to a current density of 0.194 A/cm2 and 0.086 A/cm2 (by dividing their exposed surface area in concrete: 51.5 cm2 and 463.3 cm2) respectively 10. These t